🤔 Thoughtful Claude - DeepSeek R1 Reasoning Server
@martinbowling
MCP server that enhances Claude's reasoning capabilities by integrating DeepSeek R1's advanced reasoning engine 🤔
Overview
What is Thoughtful Claude - DeepSeek R1 Reasoning Server?
An MCP server that enhances Claude’s reasoning capabilities by integrating DeepSeek R1’s advanced reasoning engine. It is designed for users who want Claude to leverage DeepSeek’s state-of-the-art reasoning model for complex multi‑step tasks.
How to use Thoughtful Claude - DeepSeek R1 Reasoning Server?
Install with Python 3.12+, the uv package manager, and a DeepSeek API key from platform.deepseek.com. Clone the repo, install dependencies (mcp[cli], httpx, python-dotenv), create a .env file with DEEPSEEK_API_KEY, then run mcp install server.py -f .env. The server auto‑starts with Claude Desktop; Claude sends queries to DeepSeek R1 for reasoning and wraps the result in <ant_thinking> tags.
Key features of Thoughtful Claude - DeepSeek R1 Reasoning Server
- Leverages DeepSeek R1’s reasoning engine
- Integrates seamlessly with Claude’s thought process
- Enterprise‑grade security via
.envand no key exposure - Full MCP protocol support with streaming responses
- Async/await architecture for efficient processing
- Proper error handling with graceful fallback messages
Use cases of Thoughtful Claude - DeepSeek R1 Reasoning Server
- Comparing numerical values (e.g., “Is 9.9 greater than 9.11?”)
- Solving logic puzzles (e.g., “If all A are B, and some B are C…”)
- Performing complex analysis (e.g., comparing quantum and classical computing)
- Answering questions that require multi‑step reasoning
FAQ from Thoughtful Claude - DeepSeek R1 Reasoning Server
What runtime requirements does the server have?
Python 3.12+ and the uv package manager are required.
How do I obtain a DeepSeek API key?
You can get one from platform.deepseek.com.
How does the server protect my API key?
The key is stored in a .env file and is never exposed in responses.
What model does the server use?
It uses the deepseek-reasoner model with streaming enabled and max tokens set to 1 (optimized for reasoning extraction).
How are errors handled?
API errors are wrapped in <ant_thinking> tags with clear messages; connection issues have a 30‑second timeout and automatic stream cleanup.
